ACIT4030 Machine Learning for 3D Computer Vision Emneplan

For the final assessment a grading scale from A to E is used, where A denotes the highest and E the lowest pass grade, and F denotes a fail.

We assume good background in programming, machine learning, and linear algebra. Knowledge of computer graphics and image processing is preferable, but not strictly required.

The thesis outline and the thesis are individual work. The thesis outline must be submitted the semester before thesis submission at the latest.

In addition to the compulsory seminars on research ethics, the students must participate in and submit work requirements for the 6 compulsory thesis seminars. The seminars require preparatory work in research methods, implementation strategies, article writing, and use of library services and search engines. Individual adaptations are dictated by the candidate’s chosen project. Students are advised to find a lab group to participate in.

Thesis supervision is compulsory. Students are assigned a field supervisor employed and trained for this purpose by the Department of Behavioral Science, with responsibility for students for the duration of the project. Field supervision is 40 hours per project, distributed according to agreement between student and supervisor. In addition, the Department of Behavioral Science provides back-up supervision as needed up to 20 hours per project. As the supervisor is formally responsible for the fulfilment of requirements by data protection authorities and ethics review boards, the thesis cannot be submitted without the written approval of the (main) supervisor.

As the thesis supervisor is formally responsible for the fulfilment of requirements by data protection authorities and ethics review boards, the thesis cannot be submitted without the written approval of the field supervisor.

The following required coursework must be approved before the student can take the exam:

a) Two research ethics seminars are 100 % attendance required before submitting the thesis outline. The seminars will take place digitally.

b) Completion of 6 attendance required, two-day digital thesis seminars with all tasks approved. These tasks are written submissions of no more than 5000 keystrokes each. The order of submissions 2 - 6 is individually determined by the student in collaboration with the field supervisor. Task approval is granted by the field supervisor.

1. An outline of a project idea, with a tentative timeline.
2. An analysis of the contingencies for implementation of the project.
3. A description and justification of search strategy and choice of background literature.
4. A justified suggestion for research design.
5. A detailed description of methods for data collection, including methods for securing data integrity, and a justified suggestion for data analysis strategies.
6. A risk and vulnerability analysis for the project, and strategies for data protection and security, including application of approval from external bodies.

c) Approval of final thesis outline by program faculty. The outline must incorporate the subjects of the tasks mentioned as Course requirements b 2 - 6.

d) Submit a signed contract with the appointed supervisor.

• Introduction to image processing and geometric modelling
• Convolutional neural networks for images and graphs
• Segmentation for images and shapes
• Correspondences and mappings
• Modelling, synthesis, and analysis
• Joint embedding for images and 3D data

Thesis. Submission of master’s thesis according to program specifications, consisting of a project report. Technical standards according to the current APA manual. Submitting dates each year: June 15th and November 15th.

Thesis format

The Master’s thesis consists of a project report. The format of the project report should conform to the publishing conventions of the science of behavior analysis, and the length should be appropriate for the work that is reported. The project report must include an executive summary for the Best Practice project bank. The project report is written according to the current APA Publishing Manual Standard; double spaced Times New Roman size 12 with standard margins for MS Office Word. Examiner’s guidelines are provided.

Ethics and data protection

Theses are submitted that require approval from the Norwegian Centre for Research Data (NSD) and or Regional Committee for Medicine and Health Sciences Research Ethics (REK). For some theses, the topic in itself will force a discussion of normative ethical evaluations, even when REK are not involved. This is the case when invasive/restrictive interventions are described; data comes from functional analysis assessment of challenging behavior, or when members of vulnerable clinical populations are subjects in the studies (this list is NOT exhaustive, and student and supervisor are jointly responsible for determining whether the ethical evaluations should be part of the main body of the thesis). In cases where advice on ethics aspects of the project is relevant, The National Committee for Research Ethics in the Social Sciences and the Humanities (NESH) should be contacted.

Web pages:

NSD https://nsd.no/

REK https://rekportalen.no/

NESH https://www.forskningsetikk.no/retningslinjer/hum-sam/forskningsetiske-retningslinjer-for-samfunnsvitenskap-humaniora-juss-og-teologi/

If the decision is made that a discussion of ethics is not required in the main body of the thesis, the student writes a reflection note - a short description of the deliberations that were made before this decision was reached (maximum 6000 keystrokes). If student and supervisor decide that there is no need for registering the project with NSD, a description of the deliberations preceding this decision is also required. Ethics and data protection can be treated in the same reflection note.

The note should describe the type of data collected, whether personal information is involved, and whether application for approval by external such as REK and NSD bodies are required. If a case number has been assigned by external bodies, it should be included.

In addition, the note should describe whether a risk and vulnerability analysis has been performed, and if it has, it should be included as an appendix.

The following required coursework must be approved before the student can take the exam:

Two mandatory group assignments consisting of technical tasks, summarized in reports (about 10 pages each).

Grade scale A-F

One internal and one external examiner will assess all exams.

For the final assessment a grading scale from A to E is used, where A denotes the highest and E the lowest pass grade, and F denotes a fail.

Two internal examiners. External examiner is used periodically.